结构突变趋势平稳过程与随机趋势过程的虚假回归研究  

Spurious Regression between Broken-Trend Stationary Process and Stochastic Trend Process

作　　者：张凌翔[1] 张晓峒[1] 

机构地区：[1]南开大学经济学院数量经济研究所

出　　处：《统计研究》2011年第5期105-110,共6页Statistical Research

基　　金：教育部项目"时间序列非平稳检验理论与应用研究"(09YJA790111)资助

摘　　要：在已有研究的基础上,本文更为深入地研究含有结构突变的趋势平稳变量与随机趋势变量间的虚假回归问题。本文推导出OLS估计下DW统计量、F统计量以及R2的极限分布,并且将回归模型扩展到动态情形下,推导出用于Granger因果检验的F统计量的极限分布;采用Monte Carlo模拟方法分析了数据生成过程的各项参数对各统计量有限样本分布的影响;最后,本文分析了在有限样本下,数据生成过程的各项参数对虚假回归及虚假Granger因果关系发生概率的影响。We study the phenomenon of spurious regression between two random variables,when the generating mechanism of one of individual series is assumed to follow a stationary process around a trend with multiple breaks in the level and slope of trend and the other is stochastic trend.We develop the relevant limit theory of the DW statistics,F statistics and R2of the OLS regression and the F statistics of Granger Causality tests.Furthermore,we analyze the distributions of these statistics in finite samples and the impact of parameters of the generating mechanism on these finite sampling distributions via Monte Carlo simulations.Finally,we present the spurious regression rates and reveal how the phenomenon of spurious regression and spurious Granger Causality relationship is to the generating mechanism parameters in finite samples.

关 键 词：结构突变 随机趋势 虚假回归 

分 类 号：O212[理学—概率论与数理统计]